The rejection of transplanted cells and tissues of allogeneic origin proves that the immune system is capable of destroying its targets. It has long been a goal of immunologists to direct the destructive capability of the immune system against a person's tumor cells and thereby effect the rejection of the tumor and the cure of the patient. Immunotherapy would be especially useful to rid a person of residual tumor cells that have spread beyond the site of the primary tumor mass. The primary tumor can usually be resected surgically or irradiated by local radiotherapy. The patient too often succumbs later, however, to metastatic tumor cells that have spread to other parts of the body. Immunotherapy would be an ideal way to destroy nests of metastatic tumor cells before they grow into large fife-threatening tumors. Lymphocytes patrol the tissues and lymphocytes sensitized to the tumor cells could kill the metastatic tumor cells remaining after resection of the primary tumor mass.
The problem for immunotherapy, however, is to activate the immune system against antigens that can mark the tumor cells for destruction. It had been hoped that tumors might bear “foreign transplantation” antigens produced by viruses or created by immunologically significant mutations of genes in the tumor cell. It now turns out that such tumor specific transplantation antigens (TSTA) are rare.
Tumor cells are characterized immunologically for the most part by TAA, which are molecules that are expressed in normal cells too. These TAA may appear normally during the early development of healthy cells, or they may be expressed normally at lower concentrations than in tumor cells. TAA may also feature minor mutations that do not appear “foreign” to the immune system and thus do not stimulate strong immune responses. TAA, for the most part, are self-antigens and, as such, they are not very immunogenic. The immune system is normally tolerant to the body's own antigens. Therefore, the induction of effective immunity against TAA is tantamount to inducing an autoimmune reaction. Immunotherapy of tumors requires the activation of the equivalent of an autoimmune reaction against the tumor cells. Moreover, it is most desirable to have the autoimmune reaction limited to the tumor itself, so the autoimmune reaction terminates once the tumor cells are destroyed. Nevertheless, autoimmunity can be compatible with life, while metastatic cancer can kill. Therefore, residued autoimmunity is a tolerable price to pay for successful tumor immunotherapy.
Examples of known TAA include p53 protein, neu differentiation factor (NDF), epidermal growth factor (EGF), carcinoembryonic antigen (CEA), and tyrosinase enzyme.
The p53 protein is the product of a tumor suppressor gene that functions to arrest the growth of mutated or aberrant cells. The p53 protein is a transcription factor that binds specifically to a consensus site present in the regulatory sequences of p53-dependent genes (el-Deiry et al., 1992; Zambetti and Levine, 1993). Mutation of the p53 gene in the domain encoding binding to the specific DNA regulatory site causes a loss of tumor suppression (Zambetti and Levine, 1993). Therefore it is not surprising that a significant proportion of natural human tumors bear mutated p53 (Hollstein et al., 1991). For reasons that are-not entirely clear, tumor cells also appear to accumulate wild-type p53 and not only mutated p53 in their cytoplasm (Moll et al., 1995). Thus the wild-type p53 molecule, and not only the mutated p53 molecule, can serve as a target for a potentially therapeutic anti-tumor immune response.
Inactivation of the p53 tumor suppressor protein by mutation of the gene or by viral insertion, gene rearrangement, or other causes is a common event in human cancers. Point mutation or deletion of the p53 gene is the most common genetic aberration in human neoplasms. Approximately 70% of colon cancers, 30 to 50% of breast cancers, 50% of lung cancers, and almost 100% of small-cell carcinomas of the lung harbor p53 mutations (Hollstein et al., 1991). The development of a tumor is often associated with accumulation in the cancer cells of the p53 protein, wild-type or mutant. Furthermore, mutated p53 proteins are tumor-specific antigens that can be recognized as targets by the immune system (Melief and Kast, 1991; Yanuck et al., 1993). Cancer patients can manifest immune responses directed to wild-type and mutant p53 proteins. The p53 protein, mutant and wild-type, can accumulate in the cytoplasm of cancer cells, and cancer patients have indeed been found to produce antibody (Lubin et al., 1993; Schlichtholz et al., 1992) and T cell responses to p53 (Houbiers et al., 1993; Tilkin et al., 1995). Normal cells express p53 to a much lower degree and, unlike tumor cells, normal cells show no accumulation of p53 in the cytoplasm. Thus, tumor cells and normal cells differ in both the amount and compartment of p53 expression. For these reasons, the wild-type p53 molecule, and not only the mutated p53 molecule, can serve as a target for a potentially therapeutic anti-tumor immune response.
To identify T-cell epitopes in p53, Houbiers et al., 1993, synthesized peptides of wild-type p53 and peptides with the point mutations of p53 detected in colorectal and ovarian cancers. Some of the p53 peptides were shown to bind in vitro to HLA-A2.1 molecules and to induce specifically cytotoxic T lymphocytes (CTL) clones. Characterization of anti-p53 immunity and its implications for tumor therapy have been studied using peptides derived from wild-type or mutated p53 sequences to elicit CTL responses in experimental animals (Noguchi et al., 1994; Noguchi et al., 1995; Yanuck et al., 1993). Mouse fibroblasts transfected with a mutated human p53 gene were specifically killed by CD8+ CTL from the spleens of mice that had been pulsed with a 21-amino acid peptide encompassing a p53 point mutation from a human lung carcinoma (Yanuck et al., 1993). A nonapeptide containing a codon 234 mutation (234CM) induced CD8+ CTL that lysed a 234CM-pulsed PIHTR mastocytoma cell line (Noguchi et al., 1994). Mice immunized with peptide 234CM were resistant to challenge with Meth A sarcoma cells (Noguchi et al., 1994), and vaccines containing peptide 234CM in the QS-21 adjuvant caused regression of established Meth A tumors in mice treated with IL-12 (Noguchi et al., 1995).
Thus, both mutated p53 and wild-type p53 are tumor-associated antigens and attempts have been made to use these molecules as immunogens for tumor immunotherapy (Houbiers et al., 1993; Noguchi et al., 1994; Noguchi et al., 1995; Yanuck et al., 1993; published PCT Application WO 94/02167). However, p53 is not very immunogenic, probably because it is a self-protein and therefore immunologically tolerated.
An antibody binds to an antigen at its variable region (antigen-binding site). Therefore, the variable regions of antibodies have three-dimensional structures that are complementary to the structures of the antigenic determinants the antibodies recognize.
The binding site of the antibody complementary to the structure of the antigen is created by hypervariable regions of the light and heavy chains of the Fab portion of the antibody. These binding site structures are formed by the collective aggregate of the CDR of the light and heavy chains of the immunoglobulin molecule (Alzari et al., 1988). However, an antibody itself, when recognized by another antibody, can be considered to be an antigen. In the case where structures of the variable regions of the antibody are recognized, these structures are called idiotypes Gd), and the antibodies that recognize the idiotypes of the antibody are called anti-idiotypic (anti-id) antibodies. The structure corresponding to the antigenic determinant of the antibody is called an idiotope (Jerne, 1974).
It has been reported that immunization with mAbs can induce immune responses that extend beyond the specificity of the antibody (Takemori et al., 1982), probably by anti-idiotypic connectivity (Jerne, 1974; Cohen, 1989, 1992) based on idiotypic determinants in the variable (V) region of the immunizing mAb (Bruggemann et al., 1980). According to idiotypic antibody network terminology, Ab1 is the first antibody, the antibody binding to the antigen, and Ab2 is the anti-idiotypic antibody to Ab1. The variable region of Ab2 may mimic the conformation of the antigen because both the antigen and Ab2 can be bound by Ab1. Ab3 is the anti-idiotypic antibody to Ab2. Because of the chain of structural complementarity, Ab1 and Ab3 can have similar specificity for the original antigen.
Antibodies have been used in the past in tumor immunotherapy in two ways: Ab1 antibodies as tumor-specific antigens on B lymphoma cells, and Ab2 antibodies as anti-idiotypic mimics of tumor antigens. Ab1 idiotypic determinants expressed by immuno globulins on the surface of neoplastic B cells have been used in experimental models as tumor-associated targets to induce protective immunity (Ab2) against B cell lymphomas which, unlike solid tumors, are particularly sensitive to antibodies (reviewed by Yefenof et al., 1993). However, Ab1 idiotypic determinants are unique to each B-cell tumor, and the practical requirements of preparing an individual protein vaccine for each patient has made the application to the clinic difficult and expensive (Stevenson et al., 1995).
Ab2 antibodies mimicking TAAs of various kinds have been used to induce antibodies (Ab3) to tumor antigens (reviewed by Wettendorf et al., 1990). However, Ab2 immunization has been usually less successful than has immunization with the TAAs themselves,(Wettendorf et al., 1990).
With regard to the possible anti-cancer effects of anti-p53) antibodies, cancer patients have been found to produce antibodies to the amino terminus of the p53 molecule, but these antibodies appear to mark the development of cancer rather than to protect against the disease (Soussi, 1996). Investigation of the effects of immunity to the central and carboxy domains of p53 might therefore be of some importance.
Published International PCT Application No. WO 94/12202 describes the activation of a mutant p53 that occurs at elevated levels in tumors and does not substantially suppress tumor growth, for specific DNA binding, wherein the mutant p53 is activated with a ligand capable of binding to, and activating the mutant p53), wherein the ligand may be the anti-p53 mAb 42l which binds to the carboxy terminal region of p53, or the bacterial heat shock protein DnaK, or a ligand which binds effectively to the same site on the mutant p53.